Emma L. Worthington1, Ben I. Moat2, David A. Smeed2, Jennifer V. Mecking2, Robert Marsh1 and Gerard D. McCarthy3
- 1 University of Southampton, European Way, Southampton, SO14 3ZH, UK
- 2National Oceanography Center, European Way, Southampton, SO14 3ZH, UK
- 3ICARUS, Department of Geography, Maynooth University, Maynooth, County Kildare, Ireland
Received: July 16, 2020 –
Discussion started: August 14, 2020 –
Revised: December 09, 2020 –
Accepted: December 21, 2020 –
Published: February 15, 2021
Between 2004 and 2012, the RAPID-MOCHA-WBTS (RAPID – Meridional Overturning Circulation and Heatflux Array – Western Boundary Time Series, hereinafter RAPID Array) observed a decrease in the strength of the Atlantic meridional overturning cycle (AMOC) with this weakened state exist until 2017. Climate model and paleo-oceanographic studies suggest that the AMOC may have declined decades or even centuries ago; However, direct observations prior to 2004 are sparse and only give “snapshots” of the overturning cycle. Previous studies have used linear models based on top layer temperature anomalies to extend AMOC estimates back in time. However, these ignore changes in deep circulation that are evident in the observations of the AMOC decline. Here we are developing an empirical model of AMOC variability with higher fidelity based on RAPID data and physically associated with changes in the thickness of the persistent upper, middle and deep water masses at 26 ° N and the associated transports. We applied historical hydrographic data to the empirical model to create an AMOC time series from 1981 to 2016. Increasing the resolution of the observed AMOC to approximately annually shows multi-year variability in line with RAPID observations and shows that the downturn between 2008 and 2012 was the weakest AMOC since the mid-1980s. However, the time series do not show a general AMOC decline as indicated by other proxies and high-resolution climate models. Our results confirm that adequately capturing changes in deep circulation is key to detecting anthropogenic AMOC decline related to climate change. Worthington, EL, Moat, BI, Smeed, DA, Mecking, JV, Marsh, R. and McCarthy, GD: A 30-year reconstruction of the Atlantic meridional overturning circulation shows no decline, Ocean Sci., 17, 285-299, https: //doi.org/10.5194/os-17-285-2021, 2021.1 Introduction
In the northern hemisphere, the Atlantic meridional overturning cycle (AMOC) transports up to 90% of the total heat transported from the subtropical Atlantic towards the poles (Johns et al., 2011), whereby the associated heat transfer to the air above is helpful for the relatively mild climate of north-western Europe maintain its latitude. The AMOC also transports fresh water to the equator and the associated deep water formation transports carbon and heat into the deep ocean (Kostov et al., 2014; Winton et al., 2013; McDonagh et al., 2015). A significant change in AMOC circulation is therefore likely to have an impact on the climate in northwestern Europe and beyond, possibly affecting global hydrological and carbon cycles. Although the Intergovernmental Panel on Climate Change (IPCC) says the AMOC is unlikely to stop this century, they state with medium confidence that a slowdown by 2050 is very likely due to anthropogenic climate change (Stocker et al., 2013) .
The importance of the AMOC means that it has been observed by the RAPID-MOCHA-WBTS mooring array (RAPID – Meridional Overturning Circulation and Heatflux Array – Western Boundary Time Series, hereinafter RAPID Array) at 26 ° N since 2004 raised the great variability of the AMOC- Transports emerged on a number of time scales (Kanzow et al., 2010; Cunningham et al., 2007), including a decrease in AMOC strength between 2004 and 2012 (Smeed et al., 2014). This reduced state continued in 2017 (Smeed et al., 2018). The decrease is more an internal variability than a long-term decrease in the response to anthropogenic forcing (Roberts et al., 2014), which is currently too short to record in the time series. Although the AMOC has been well monitored at 26 ° N since 2004, AMOC strength estimates previously were limited to cases of transatlantic hydrographic stretches along 24.5 ° N in 1957, 1981, 1992, 1998, and 2004, which only foresaw Snapshots of the overturning circulation strength (Bryden et al., 2005). Extensive additional hydrographic data around 26 ° N are available, particularly on the western border, which, however, are not sufficient to reconstruct the AMOC conventionally (Longworth et al., 2011). Due to the limited availability of hydrographic data, proxies were used to reconstruct the AMOC time series prior to 2004.
In a proxy reconstruction, Frajka-Williams (2015) used the sea surface height from satellite altimetry to estimate the baroclinic trans-basin transport between 1993 and 2014 at 26 ° N. In another case, Longworth et al. (2011) used a temperature anomaly at the western border as a proxy for geostrophic transport within the upper 800 m or the thermocline. The temperature anomaly at 400 dbar explained 53% of the variance in thermocline transport. Both Longworth et al. (2011) and Frajka-Williams (2015) used single-layer models that do not take into account the variable depth structure of the AMOC in the subtropics.
At 26 ° N, the dynamics of the AMOC include several water masses flowing in different layers in opposite directions, driven by the density structure that changes with depth (Fig. 1a). Within the permanent thermocline, which is up to 800 m deep on the western border and up to 600 m on the eastern border, isopyknals rise towards the eastern border, indicating a current to the south (Hernández-Guerra et al., 2014). Below the thermocline, the isopyknalen deepen to the east, and the resulting transport profile (Fig. 1c) shows a small transport to the north, which is centered around 1000 m and lies above and below between the transports to the south. Although called Antarctic Intermediate Water (AAIW) by RAPID, both AAIW and Mediterranean water are observed between 700 and 1600 m on the eastern border, with the relative contribution varying seasonally in each case (Fraile-Nuez et al., 2010; Machín and Pelegrí, 2009 ; Hernández-Guerra et al., 2003). The transport profile also shows the North Atlantic Deep Water (NADW), which has two distinct layers: Upper (UNADW) above 3000 m, which is mainly formed in the Labrador Sea (Talley and McCartney, 1982), and Lower (LNADW) below 3000 m has its origin in the overflows from the North Sea (Pickart et al., 2003). Changes observed in one NADW layer are not necessarily observed in another. Smeed et al. (2014) found that the decrease in AMOC strength between 2004 and 2012 was observed in LNADW but not in UNADW, while Bryden et al. (2005) found that the LNADW transport estimated from transatlantic hydrographic sections at 25 ° N decreased from -15 Sv in 1957 to less than -7 Sv in 1998 and 2004, while the UNADW transport decreased between -9 and –12 Sv remained. Below the NADW layers is a small transport to the north below 5000 m, Antarctic Bottom Water (AABW), which flows along the west side of the mid-Atlantic ridge. The division between the upper south and deep south transport defines the strength of the overturning cycle: a weak AMOC is associated with a stronger recirculation within the upper layers of the thermocline and a weaker deep reflux; A stronger AMOC is associated with a weaker thermocline recirculation and a stronger deep NADW transport. In order for an empirical model to be able to more fully represent the AMOC dynamics, especially changes with lower frequency, we propose that it must represent these deeper layers. A layer model interpretation of the density structure and the associated water mass transport is shown in Fig. 1b.
Figure 1 (a) World Ocean Circulation Experiment (WOCE) North Atlantic section A05 with neutral density γn (kg m – 3) at 24 ° N, July or August 1992. From the WOCE Atlantic Atlas Vol. 3. (Koltermann et al., 2011). (b) Scheme of four dynamic layers that are to be represented within the regression model by density anomalies on the western and eastern borders at a depth within each layer. The density anomalies are represented by the circular markers. (c) Profile of the RAPID-estimated mean transport in the middle of the ocean and the resulting stratified transports to the north and south. The mean AMOC depth is approx. 1100 m.
Here we rethink the Longworth et al. (2011) by using linear regression models to represent the AMOC and to further develop the method to include additional layers that are representative of deep circulation. Section 2 describes how we trained and validated our statistical model using the RAPID dataset and how we selected historical hydrographic data to apply to the model. Section 3 describes how this hydrographic data was used to create an expanded time series of AMOC strength from 1982 to 2016. Insects. In Figures 4 and 5, we discuss the implications of establishing the longest observation time series of AMOC strength that takes into account the variability in the deep NADW layers and recognize the limits of using an empirical model.
Read the full paper here.